OpenAI: GPT-5.6 Luna Pro

GPT-5.6 Luna Pro is a text model from OpenAI, identified on api14 as openai/gpt-5.6-luna-pro. It shares its underlying model with GPT-5.6 Luna, but it is served with reasoning.mode set to pro, which is intended to produce higher-quality responses on complex tasks. The provider release date is 9 July 2026, and the knowledge cutoff is 16 February 2026.

The model accepts text, images and files as input, and it returns text. Its context window is 1,050,000 tokens, with a maximum output of 128,000 tokens. Supported request parameters include include_reasoning, max_completion_tokens, max_tokens, reasoning, reasoning_effort, response_format, seed, structured_outputs, tool_choice, tools, and verbosity. Because reasoning mode is part of how this variant is served, the reasoning and reasoning_effort parameters are relevant when you want to steer how much deliberation the model applies.

Those capabilities point to work where the input is large and mixed. A long contract, a technical manual, or a repository of documentation can be supplied alongside images or files, and the model can produce a written answer in one pass rather than through chunking. Tool calling and structured outputs suit pipelines where the model must fetch data or return JSON that downstream code parses. The seed parameter helps when you need reproducible sampling, and response_format gives control over the shape of the returned text. Reasoning controls are useful for tasks that reward deliberate analysis, such as multi-step planning, code review, or reconciling conflicting information across a large document set.

To call the model through api14, use the OpenAI-compatible chat completions endpoint with the model id openai/gpt-5.6-luna-pro. Existing OpenAI client libraries work without changes beyond the base URL and API key. Send messages with text, image, or file content as needed, and include any of the supported parameters in the request body. If you request reasoning output, the include_reasoning parameter controls whether that content is returned. Note that pro mode spends far more than the standard Luna variant, so it is worth reserving for requests that genuinely need the extra deliberation. For lighter or high-volume traffic, the standard GPT-5.6 Luna model with reasoning.mode left at its default is the more economical choice.